论文信息 - Condensed List Relevance Models

Condensed List Relevance Models

Pseudo-relevance feedback has traditionally been implemented as an expensive re-retrieval of documents from the target corpus. In this work, we demonstrate that, for high precision metrics, re-ranking the original feedback set provides nearly identical performance to re-retrieval with significantly lower latency.

Fernando Diaz | Fernando Diaz

[1] W. Bruce Croft,et al. Latent concept expansion using markov random fields , 2007, SIGIR.

[2] W. Bruce Croft,et al. Effective query formulation with multiple information sources , 2012, WSDM '12.

[3] Hao Wu,et al. An incremental approach to efficient pseudo-relevance feedback , 2013, SIGIR.

[4] Fernando Diaz,et al. Regularizing query-based retrieval scores , 2007, Information Retrieval.

[5] Craig MacDonald,et al. Effect of dynamic pruning safety on learning to rank effectiveness , 2012, SIGIR '12.

[6] James Allan,et al. Fast query expansion using approximations of relevance models , 2010, CIKM.

[7] Rong Jin,et al. Web image retrieval re-ranking with relevance model , 2003, Proceedings IEEE/WIC International Conference on Web Intelligence (WI 2003).

[8] Aviezri S. Fraenkel,et al. Local Feedback in Full-Text Retrieval Systems , 1977, JACM.

[9] W. Bruce Croft,et al. Using Probabilistic Models of Document Retrieval without Relevance Information , 1979, J. Documentation.

[10] Robert Wing Pong Luk,et al. A Generative Theory of Relevance , 2008, The Information Retrieval Series.

[11] Andrei Z. Broder,et al. Efficient query evaluation using a two-level retrieval process , 2003, CIKM '03.

[12] Tetsuya Sakai,et al. Alternatives to Bpref , 2007, SIGIR.